Tail command limitations?

[James Daniel]

Does the "tail -f output.log" command in Unix always capture all output or does it pare it down if too much data is being outputted? Example: If I start tail -f command, and 10 seconds later an external program outputs 1000 lines of log file in period of 1 second or less, will the output contain all 1000 lines - or will it be truncated in some way? The reason why we want tail to be 100% accurate is because we plan on using the output of tail in a java program to process the logs automatically and we don�t want to lose any information. I was under the assuption that tail always captured all output no matter the rate of output. Is this incorrect?

[ April 23, 2007: Message edited by: James Daniel ]
[ April 23, 2007: Message edited by: James Daniel ]

[Ernest Friedman-Hill]

Is is not incorrect -- i.e., tail won't drop any data.

[James Daniel]

Ernest Friedman-Hill,

Thanks for the prompt reply. That is what we thought. A developer in our group was telling us this and we needed to verify. Thanks again..

[Tim Holloway]

Tail does a seek to the end of the file, which is a basic function in Unix. So it's actually quite fast.